OpenCPN Beta Version 3.3.1824 Released

[bcn]

Dave,

no success with the AIS list bug (ARMhf).
And it's even simpler: just toggling one Track to YES and then closing with x will crash.
SHOW ALL TRACKS and x is ok.

Hubert

[CarCode]

Originally Posted by bdbcat

Hi there...

I have just posted a new commit to github which addresses the AIS target list closure crash bug.

Please have a look if you can build from source.

Thanks
Dave

Dave,
there is a semantic issue at

src/AIS_Decoder.cpp:1081:

if( dse_mmsi || b_take_dsc){
        if( m_ptentative_dsctarget ){

Variable 'pTargetData' is used uninitialized whenever 'if' conditions are false.


Gerhard

[bdbcat]

Hubert...

Any chance of a stack trace on the ARMhf crash?

I have never actually seen a crash on my various systems on closing the list, so I am working blind.

Dave

[bcn]

Dave...

a HowTo/hint for beginners under Debian? I'm using just a script for cake/built.
In the log there is nothing of course.

Sean perhaps should be able to test this as well. The basic platform of the Cubies is identical to the Olimex (A20 processor).

Hubert

[bdbcat]

Hubert...

I think gdb and/or ddd are available for armhf.

So, try this:
A. with ddd
$sudo apt-get install ddd

$ddd opencpn

Program->Run

Make it crash

Status->Backtrace

Take a screenshot,

or

B. With gdb

gdb opencpn

(gdb) r

Make it crash

(gdb) bt

Screenshot, or cut/paste the output.

Good Luck
Dave

[bdbcat]

Gerhard...

Good catch.

Fixed in github now.

Thanks
Dave

[bcn]

Originally Posted by bdbcat

Hubert...

So, try this:
A. with ddd
$sudo apt-get install ddd

$ddd opencpn

Program->Run

Make it crash

Status->Backtrace

Take a screenshot,

Good Luck
Dave

here we go.
The debug info:

Starting program: /usr/local/bin/opencpn 
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/arm-linux-gnueabihf/libthread_db.so.1".
[New Thread 0xb12c52a0 (LWP 3137)]
[New Thread 0xb0ac52a0 (LWP 3138)]
Program received signal SIGSEGV, Segmentation fault.
0xb6e3c85c in wxWindowBase::Close(bool) ()
   from /usr/lib/arm-linux-gnueabihf/libwx_gtk2u_core-2.8.so.0
(gdb) bt
#0  0xb6e3c85c in wxWindowBase::Close(bool) ()
   from /usr/lib/arm-linux-gnueabihf/libwx_gtk2u_core-2.8.so.0
#1  0x0018cb88 in AISTargetListDialog::OnPaneClose (this=0xae5f70, event=...)
    at /root/GitOpenCPN/OpenCPN/src/AISTargetListDialog.cpp:633
#2  0xb6f2fbd4 in wxAppConsole::HandleEvent(wxEvtHandler*, void (wxEvtHandler::*)(wxEvent&), wxEvent&) const ()
   from /usr/lib/arm-linux-gnueabihf/libwx_baseu-2.8.so.0
#3  0xb6f7e99c in wxEvtHandler::ProcessEventIfMatches(wxEventTableEntryBase const&, wxEvtHandler*, wxEvent&) ()
   from /usr/lib/arm-linux-gnueabihf/libwx_baseu-2.8.so.0
#4  0xb6f7ec62 in wxEvtHandler::SearchDynamicEventTable(wxEvent&) ()
   from /usr/lib/arm-linux-gnueabihf/libwx_baseu-2.8.so.0
#5  0xb6f7ecd2 in wxEvtHandler::ProcessEvent(wxEvent&) ()
   from /usr/lib/arm-linux-gnueabihf/libwx_baseu-2.8.so.0
#6  0xb6b14818 in wxAuiManager::ProcessMgrEvent(wxAuiManagerEvent&) ()
   from /usr/lib/arm-linux-gnueabihf/libwx_gtk2u_aui-2.8.so.0
#7  0xb6b181d8 in wxAuiManager::OnFloatingPaneClosed(wxWindow*, wxCloseEvent&)
    () from /usr/lib/arm-linux-gnueabihf/libwx_gtk2u_aui-2.8.so.0
#8  0xb6b1c15c in wxAuiFloatingFrame::OnClose(wxCloseEvent&) ()
   from /usr/lib/arm-linux-gnueabihf/libwx_gtk2u_aui-2.8.so.0
#9  0xb6f2fbd4 in wxAppConsole::HandleEvent(wxEvtHandler*, void (wxEvtHandler::*)(wxEvent&), wxEvent&) const ()
   from /usr/lib/arm-linux-gnueabihf/libwx_baseu-2.8.so.0

Hubert

[bcn]

Dave,

ok now with the latest Git pull.

Hubert

[seandepagnier]

<message deleted>

[CarCode]

Originally Posted by bcn

Dave,

ok now with the latest Git pull.

Hubert

So I guess I was correct with my assumption about the uninitialized variable above. In Xcode I got a warning about this but other compilers might be silence and one often search in the wrong place to fix something in such cases.

Gerhard

[Ptizef]

Hi Dvae

With the last git compiled, it seems that crashes are behind us now
Just a curious behavior with the AIS alarm dialog which is uncompleted
The acknowledge button is missing. I never noticed that before
see screen shot attached
Jean Pierre

[rgleason]

Jeez you guys are ggoodd. How to do a backtrace in Win?

[bdbcat]

jp...

Thanks for the catch on AIS Alert dialog bug.

Simple typing error, too late at night.
Fixed in github now.

Dave

[CarCode]

Is there a kind soul who can quality control the code a little bit before the final version 3.4 is released? It seems to be mostly dead values of variables, null pointer problems and the such, see analytic results attached (remove .doc).

Gerhard

[NEVERMIND]

Hi,
With the last version and XP when you mouve the chart there is a background black at the corner and after Opencpn shut down.
Jean Paul